All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=864970

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=507e45f7c2c8dc216c01f111bf9683874802c5d62560ba4a24bd243ba823f230


Bot(s) for this bug's original alert(s):

linux-perf

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16aedf05a40000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/16aedf05a40000

Roll fontconfig to d1f48f11 by thomasanderson@chromium.org
https://chromium.googlesource.com/chromium/src/+/3ff47e60c5ce645973e57fc8dbe88b43d63edaf4
7.722 → 8.182 (+0.46)

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Original fontconfig roll introduced a bug:
https://chromium.googlesource.com/chromium/src/+/a0c1584a2fade0146b7cd3380abe42feda02ad7c

Subsequent fontconfig roll fixed the regression introduced in the previous roll:
https://chromium.googlesource.com/chromium/src/+/a0c1584a2fade0146b7cd3380abe42feda02ad7c

The graphs indicate a performance improvement on the original roll that introduced the bug, and a regression after.  These performance differences are expected due to the nature of the bug introduced and then fixed.